Sensory information coding in frog semicircular canals: how spike firing depends on EPSP rate

The frequency response of the cytoneural junction was studied in the frog labyrinth posterior canal by recording EPSPs in response to sinusoidal rotation at various frequencies (0.02-1 Hz) in the presence of tetrodotoxin. Peak accelerations ranged from 3 to 63 deg/s2 but were kept constant (15 deg/s2 ) in some units by changing rotation amplítudes and frequencies. EPSP waveform was evaluated from power spectra and EPSP rate was determined by means of a fluctuation analysis procedure devised to study re¬sponses as fast as the changes in EPSP rate during rotation at 1 Hz. At all rotation frequencies the responses were asymmetric, the increase in EPSP rate for excitatory acceleration being 1.2-2.5 larger than the decrease during the corresponding inhibitory acceleration. Asymmetrical sinusoids well fit all responses. Frequency responses were obtained for each unit in terms of: 1)GAIN (change in EPSP rate vs. acceleration);. 2)PHASE (phase lead vs. angular velocity); 3)ASYMMETRY (ratio between positive and negative responses). At constant acceleration the relation of gain vs. frequency was described by a low-pass filter function (corner frequency range 0.05-0.14 Hz). This is in agreement with the behaviour predicted by the cupula-endolymph system. When peak accelerations were changed the responses, normalized to peak acceleration, displayed the same behaviour (corner frequency 0.05¬0.19 Hz); phase leads were generally in agreement with the model. Asymmetry in the response did not exhibít any clear-cut frequency dependence. In conclusion, these data indicate that most of the static and dynamic properties usually described for the spike discharge pattern directly derive from the properties of neurotransmitter release at the cytoneural junction